Gibbs sampling for parsimonious Markov models with latent variables
Ralf Eggeling, Pierre-Yves Bourguignon, Ivo Große, Martin Luther · 2008
We propose a Bayesian model averaging approach for learning mixtures of parsimonious Markov models that is based on Gibbs sampling. The challenging problem is sampling one out of a large number of model structures. We solve it by an ecient dynamic programming algorithm. We apply the resulting Gibbs sampling algorithm to splice site classication (2), an important problem from computational biology, and nd the Bayesian approach to be superior to the non-Bayesian classication. PCTs